<![CDATA[The increasing use of artificial intelligence (AI) in language education has raised interest in the characteristics of AI-generated feedback. While previous studies have focused on the effectiveness of ChatGPT in improving writing, limited attention has been given to its interpersonal and pragmatic dimensions. Drawing on Brown and Levinson’s politeness theory, this study investigates the politeness strategies employed by ChatGPT in providing corrective feedback on EFL learners’ writing. Using qualitative content analysis, 150 feedback instances generated by ChatGPT-4 in response to essays written by 30 Indonesian undergraduate EFL students were analyzed. The findings indicate that negative politeness was the most frequently employed strategy, characterized by hedges, modal verbs, and indirect suggestions. Positive politeness and bald on-record strategies were also identified, whereas off-record strategies were absent. The results suggest that ChatGPT tends to deliver corrective feedback in a face-sensitive manner by framing corrections as suggestions rather than directives. The study highlights the importance of considering the pragmatic aspects of AI-generated feedback and contributes to the growing discussion of AI-assisted language learning.]]>
Copyrights © 2026
